✨ New floating point number arbitraries

[dubzzz]

In a nutshell

Until now fc.float was able to generate random floating numbers. Unfortunately it comes with multiple drawbacks:

• it cannot generate all the existing floats but just a limited subset
• its default is problematic: we ask for floats we get values in [ 0 ; 1 [
• it can produce non 32-bit values when asked for custom ranges

This new implementation is not anymore uniform on [ 0 ; 1 [ but it can:

• generate any 32 bits floating point numbers
• all the existing floats have the same probability to be selected (when not biased)
• infinity, nan and extreme values are activated by default

Related to #992

✔️ New feature
❌ Fix an issue
❌ Documentation improvement
❌ Other: tests

(✔️: yes, ❌: no)

Potential impacts

None

[codesandbox-ci]

This pull request is automatically built and testable in CodeSandbox.

To see build info of the built libraries, click here or the icon next to each commit SHA.

Latest deployment of this branch, based on commit 91b8b83:

Sandbox	Source
Vanilla	Configuration
fast-check-examples	Configuration

[coveralls]

Coverage increased (+0.04%) to 96.353% when pulling 91b8b83 on feat/float-next into c1d6b5a on master.